Imatinib mesylate has shown remarkable promise in the treatment of chronic myeloid leukemia (CML); however, in some patients it is clearly insufficient as a single agent to produce long-term disease-free or overall survival. AIIogeneic hematopoietic cell transplant-based therapies remain an established curative approach, but complications such as graft-versus-host disease and the need for a suitable donor limit its application. The use of autografts to support high-dose chemotherapy was being explored extensively prior to the introduction of imatinib and promising results with both in vitro and in vivo-purged autologous cells were reported. In the interim, our understanding of the abnormal biology of CML stem cells has advanced significantly and methods for quantitating different types of transplantable normal and leukemic stem cells and for expanding normal cells with rapid, short-term reconstituting potential have been devised. These advances provide an attractive basis for designing an improved strategy for purging CML stem cells from patients' autografts for use with myeloablative chemotherapy in patients without other therapeutic options. Here we will evaluate three purging methods to selectively eliminate CML stem cells: ex vivo pharmacologic treatment with Gleevec and mafosfamide; a novel culture-purging method that exploits the known intrinsically determined, highly defective self-renewal property of CML stem cells under conditions that are predicted to simultaneously amplify residual normal short-term repopulating cells typically present in many CML patients; and a combination of these. We plan to proceed as rapidly as possible to a clinical test of these approaches, in which we will compare the results obtained when patients are randomly assigned to one of the three purging methods. Key to the success of this project is the commitment of two leading groups with complementary expertise in the biology of normal and CML stem cells and the development and clinical evaluation of cultured autografts. The proposed work will thus build on an exciting history in these areas in an attempt to address an emerging need for innovative and broadly applicable approaches to the treatment of CML - prompted by a growing recognition of the limitations of imatinib mesylate as a single therapeutic agent.